Determining how spontaneous seizures initiate in patients with temporal lobe epilepsy is our long-term goal. The main hypothesis of this project is that a structural or neurochemical abnormality, usually within the ventral hippocampal formation, causes occasional, spontaneous, and progressive increases in neuronal activity, and when a threshold is passed a seizure initiates. The hypothesis predicts that a structural or neurochemical abnormality would be most severe in seizure-initiating regions that increased pre-ictal activity of principal neurons would be earliest in seizure-initiating regions, and that fcal inhibition of seizure-initiating regions would reduce seizure frequency. We propose to use epileptic pilocarpine-treated rats to test these predictions in the following aims. Specific Aim 1 s to test whether GABAergic neuron loss is most severe in seizure- initiating regions. Rats will be implanted with fine wires bilaterally in the ventral hippocampus, ventral subiculum, dorsal hippocampus, entorhinal cortex, amygdala, olfactory cortex, and septum, which are brain regions that have displayed earliest spontaneous seizure activity in the rat model. In individual subjects relative timing of electrographic onsets of spontaneous seizures will be measured to identify seizure onset site(s). In situ hybridization for glutamic acid decarboxylase and stereology will be used to quantify GABAergic neuron loss and determine whether it is more severe in seizure-initiating regions. Specific Aim 2 is to test whether action potential firing rats of principal neurons in seizure-initiating regions increase earliest before electrographic onsets o spontaneous seizures. Fine wires will be used to obtain local field potential recordings bilaterall from brain regions previously found to have earliest spontaneous seizure activity. In the same animals, tetrodes will be used to obtain single-unit recordings from principal neurons in the ventral subiculum. Spontaneous seizures will be sorted according to whether they initiated in the ventral subiculum versus elsewhere, and the timing of significantly increased average preictal firing rates will be compared. Specific Aim 3 is to test whether focal inhibition at electrographic onset sites is necessary and sufficient to block spontaneous seizures. Will inhibition of any preictally activated node in a network block ictogenesis? Or, is it necessary to inhibit the precis site of electrographic seizure onset? To address these questions, fine wires will be used to obtain local field potential recordings bilaterally from brain regions previously found to have earliest spontaneous seizure activity. A cannula will be implanted and a mini-osmotic pump will be used to continuously deliver vehicle for 1 month then tetrodotoxin for 1 month to focally inhibit activity in the ventral subiculum. Seizure frequency/severity/duration will be compared during vehicle- versus tetrodoxin-infusion. Effects of inactivating the ventral subiculum will be compared with respect to seizure onset site (ventral subiculum versus elsewhere). The proposed experiments address the most fundamental question in epilepsy research: How do spontaneous seizures start?